The validation of toxicological prediction models

An alternative method is shown to consist of two parts: the test system itself; and a prediction model for converting in vitro endpoints into predictions of in vivo toxicity. For the alternative method to be relevant and reliable, it is important that its prediction model component is of high predictive power and is sufficiently robust against sources of data variability. In other words, the prediction model must be subjected to criticism, leading successful models to the state of confirmation. It is shown that there are certain circumstances in which a new prediction model may be introduced without the necessity to generate new test system data.